search

EBISPOT / goci

GWAS Catalog Ontology and Curation Infrastructure
Apache License 2.0
26 stars 19 forks source link

Investigate unexpected sumstats+metadata submission by external user #1311

Closed eks-ebi closed 4 months ago

eks-ebi commented 5 months ago

An external user seems to have been able to create a sumstats + metadata submission for a publication (https://www.ebi.ac.uk/gwas/depo-curation/submissions/661e2936d222f50001984929), even though there was already curated data for the publication in the curation database, and a metadata-only submission already existed (https://www.ebi.ac.uk/gwas/depo-curation/submissions/62823f4e28d70900018e63e1). This should not occur, because the presence of an existing metadata submission should cause a study to appear as CLOSED FOR SUBMISSION and therefore prevent any new metadata submissions.

This is a possible loop-hole in the process, which needs further investigation.

Relevant background:

  1. The original metadata-only submission was created by Lizzy in 2022.
  2. The author later made a sumstats-only submission and attempted to add additional studies to the pre-filled template, which is impossible, therefore their submission was INVALID.
  3. The author emailed to ask us to add the additional studies to the Catalog.
  4. I (@eks-ebi) unpublished the original curated studies using the curl command, in order to edit the metadata-only submission and add the new requested studies.
  5. We deleted the invalid sumstats-only submission in order to allow the changes to the metadata-only submission to be imported into the curation app: see goci #1286
  6. I could not import the updated metadata-only submission at first, and asked @sajo-ebi to manually change the publication status, in order to enable the import. I'm not sure from the notes in goci #1286 whether any action was done at this point, or whether the import was re-enabled simply by waiting overnight.
  7. The next day, I was able to import the updated metadata-only submission.
  8. It seems that at some point after the status was changed (if it was) and before I actually imported the updated metadata-only submission, the author created the new sumstats + metadata submission.
ljwh2 commented 4 months ago

@sajo-ebi will add details of investigation

eks-ebi commented 4 months ago

@sajo-ebi We discussed this morning that since there is no clear explanation for this situation, we can go ahead and clean up the two submissions. I have now saved a copy of the sumstats, so I am ready for you to delete the author’s submission: 661e2936d222f50001984929.

sajo-ebi commented 4 months ago

@eks-ebi the submission '661e2936d222f50001984929' has been deleted

eks-ebi commented 4 months ago

Thanks @sajo-ebi !

eks-ebi commented 4 months ago

I'll leave the ticket open for you to add any details from the investigation

eks-ebi commented 4 months ago

I've now made a successful sumstats-only submission for this publication, so it should be finished now.

sajo-ebi commented 4 months ago

I investigated the DB & code couldn't find any scenario where the user can create two submissions, the logs are not there to determine what exactly happened